2021
DOI: 10.1017/psrm.2021.57
|View full text |Cite
|
Sign up to set email alerts
|

The micro-task market for lemons: data quality on Amazon's Mechanical Turk

Abstract: While Amazon's Mechanical Turk (MTurk) has reduced the cost of collecting original data, in 2018, researchers noted the potential existence of a large number of bad actors on the platform. To evaluate data quality on MTurk, we fielded three surveys between 2018 and 2020. While we find no evidence of a “bot epidemic,” significant portions of the data—between 25 and 35 percent—are of dubious quality. While the number of IP addresses that completed the survey multiple times or circumvented location requirements f… Show more

Help me understand this report

Search citation statements

Order By: Relevance

Paper Sections

Select...
2
1
1

Citation Types

4
24
0

Year Published

2021
2021
2024
2024

Publication Types

Select...
10

Relationship

0
10

Authors

Journals

citations
Cited by 66 publications
(28 citation statements)
references
References 33 publications
4
24
0
Order By: Relevance
“…These criteria resulted in exclusion of 321 (28.6%) of respondents in one of the experimental conditions, consisting of 65 excluded for both reasons, 184 excluded for the attention check alone, and 62 excluded on the basis of time on task alone. This exclusion rate is consistent with estimates of untrustworthy MTurk workers [ 34 ]. The resulting sample consists of n = 270 for the Male-Narr -Video group, n = 255 for the Female-Narr -Video group, n = 282 for the text group, and n = 393 for the control group.…”
Section: Methodssupporting
confidence: 89%
“…These criteria resulted in exclusion of 321 (28.6%) of respondents in one of the experimental conditions, consisting of 65 excluded for both reasons, 184 excluded for the attention check alone, and 62 excluded on the basis of time on task alone. This exclusion rate is consistent with estimates of untrustworthy MTurk workers [ 34 ]. The resulting sample consists of n = 270 for the Male-Narr -Video group, n = 255 for the Female-Narr -Video group, n = 282 for the text group, and n = 393 for the control group.…”
Section: Methodssupporting
confidence: 89%
“…The current study utilizes a sample of participants obtained via Amazon Mechanical Turk (MTurk). Previous work has illustrated the imperfect nature of MTurk samples (e.g., Ahler, Roush, & Sood, 2019), and potential misuse of the platforms (e.g., by “bots”) is an important issue not to be overlooked. In the current study, the description of the HIT did not mention aging, though we do note that including IP address monitoring and/or geotagging in future IRB protocols can help identify those who respond untruthfully.…”
Section: Discussionmentioning
confidence: 99%
“…Armitage and Eerola (2020) have shown that the results of music cognition experiments on chord perception carried out in standard lab settings are comparable with those from online experiments that recruit participants using services similar to MTurk. However, study of crowdsourcing platforms has shown that the percentage of non-serious participants and survey bots is relatively large (Ahler et al, 2019;Dennis et al, 2020). Hence, a pre-test and other ways to detect non-serious participants and survey bots were included in our experiment (see Participants and below).…”
Section: Methodsmentioning
confidence: 99%